EVChargeEnv / openenv.yaml
oozan's picture
Upload folder using huggingface_hub
a283aa6 verified
id: EVChargeEnv-v0
name: EVChargeEnv
version: "0.1.0"
description: >
EVChargeEnv is a continuous-control electric vehicle charging environment
with dynamic pricing, fluctuating grid load, and multi-objective reward signals.
It is suitable for benchmarking agentic behavior and testing adaptation
to non-stationary conditions.
authors:
- name: Ozan Özayranci
github: "https://github.com/oozan"
license: mit
environment:
observation_space:
shape: [4]
type: box
description:
- charge_level (0–1)
- price (0–1)
- grid_load (0–1)
- time_step_norm (0–1)
action_space:
shape: [1]
type: box
description: continuous charge rate (0–1)
reward_components:
- progress_reward
- cost_penalty
- overload_penalty
- time_penalty
termination_conditions:
- charge >= 1.0
- max_steps reached
scenarios:
- easy
- medium
- hard
entry_point: env.ev_charge_env:EVChargeEnv
tags:
- energy
- control
- continuous
- stochastic
- reinforcement-learning
- openenv